Denial of service in nghttp2
In nghttp2 before version 1.41.0, the overly large HTTP/2 SETTINGS frame payload causes denial of service. The proof of concept attack involves a malicious client constructing a SETTINGS frame with a length of 14,400 bytes (2400 individual settings entries) over and over again. The attack causes the CPU to spike at 100%. nghttp2 v1.41.0 fixes this vulnerability. There is a workaround to this vulnerability. Implement nghttp2_on_frame_recv_callback callback, and if received frame is SETTINGS frame and the number of settings entries are large (e.g., > 32), then drop the connection.
02 / AFFECTED SOFTWARE
Affected packages
195 explicit affected versions
03 / CONNECTIONS
Connected vulnerabilities
04 / EVIDENCE
Source records
In nghttp2 before version 1.41.0, the overly large HTTP/2 SETTINGS frame payload causes denial of service. The proof of concept attack involves a malicious client constructing a SETTINGS frame with a length of 14,400 bytes (2400 individual settings entries) over and over again. The attack causes the CPU to spike at 100%. nghttp2 v1.41.0 fixes this vulnerability. There is a workaround to this vulnerability. Implement nghttp2_on_frame_recv_callback callback, and if received frame is SETTINGS frame and the number of settings entries are large (e.g., > 32), then drop the connection.
In nghttp2 before version 1.41.0, the overly large HTTP/2 SETTINGS frame payload causes denial of service. The proof of concept attack involves a malicious client constructing a SETTINGS frame with a length of 14,400 bytes (2400 individual settings entries) over and over again. The attack causes the CPU to spike at 100%. nghttp2 v1.41.0 fixes this vulnerability. There is a workaround to this vulnerability. Implement nghttp2_on_frame_recv_callback callback, and if received frame is SETTINGS frame and the number of settings entries are large (e.g., > 32), then drop the connection.
In nghttp2 before version 1.41.0, the overly large HTTP/2 SETTINGS frame payload causes denial of service. The proof of concept attack involves a malicious client constructing a SETTINGS frame with a length of 14,400 bytes (2400 individual settings entries) over and over again. The attack causes the CPU to spike at 100%. nghttp2 v1.41.0 fixes this vulnerability. There is a workaround to this vulnerability. Implement nghttp2_on_frame_recv_callback callback, and if received frame is SETTINGS frame and the number of settings entries are large (e.g., > 32), then drop the connection.
05 / REFERENCES
Further evidence
- http://lists.opensuse.org/opensuse-security-announce/2020-06/msg00024.html
- https://github.com/nghttp2/nghttp2/commit/336a98feb0d56b9ac54e12736b18785c27f75090
- https://github.com/nghttp2/nghttp2/commit/f8da73bd042f810f34d19f9eae02b46d870af394
- https://github.com/nghttp2/nghttp2/security/advisories/GHSA-q5wr-xfw9-q7xr
- https://lists.debian.org/debian-lts-announce/2021/10/msg00011.html
- https://lists.debian.org/debian-lts-announce/2023/10/msg00023.html
- https://lists.fedoraproject.org/archives/list/package-announce%40lists.fedoraproject.org/message/4OOYAMJVLLCLXDTHW3V5UXNULZBBK4O6/
- https://lists.fedoraproject.org/archives/list/package-announce%40lists.fedoraproject.org/message/AAC2AA36OTRHKSVM5OV7TTVB3CZIGEFL/
- https://www.debian.org/security/2020/dsa-4696
- https://www.oracle.com//security-alerts/cpujul2021.html
- https://www.oracle.com/security-alerts/cpuapr2022.html
- https://www.oracle.com/security-alerts/cpujan2021.html
- https://www.oracle.com/security-alerts/cpujul2020.html
- https://www.oracle.com/security-alerts/cpuoct2020.html
- https://nvd.nist.gov/vuln/detail/CVE-2020-11080